Graduate Intern: LLM Reliability & Uncertainty for AI Science

National Renewable Energy Laboratory; Colorado; On-site; USD 44,500 - 71,200; Part time

Benefits offered by this job

Medical Insurance

403(b) Employer Match

Sick leave

Annual salary range

Job summary

The National Renewable Energy Laboratory is offering a graduate internship focused on quantifying uncertainty and reducing hallucinations in LLM-based science assistants. You will explore multi-turn, agentic workflows, assess internal representations, and develop detectors and probing methods, with experiments on HPC systems.

You will contribute to documentation, reports, and presentations while building robust research code and evaluation pipelines under a multidisciplinary team in Golden,

Qualifications

• Minimum 3.0 cumulative GPA.
• Undergraduate: Full‑time student in a bachelor’s degree program; or earned a bachelor’s degree within the past 12 months.
• Graduate: Full‑time student in a master’s degree program; or earned a master’s degree within the past 12 months.
• Graduate + PhD: Completed master’s degree and enrolled as a PhD student.
• Eligible for an internship period of up to one year.
• Applicants must submit official or unofficial transcripts and a letter of recommendation (if selected).

Responsibilities

• Research and evaluate uncertainty quantification and hallucination detection methods for multi‑turn, agentic scientific workflows.
• Develop probing methods that predict, from a model’s internal representations, when a scientific task specification is incomplete or inconsistent and a clarifying question is warranted.
• Build and instrument evaluation pipelines that capture and analyse model internal states over multi‑turn scientific dialogue on HPC systems.
• Conduct experiments and analyse model behaviour across computational science domains and established benchmarks.
• Contribute to technical documentation, research reports, publications, and presentations summarising project progress and findings.
• Develop, test, and maintain high‑quality research code and evaluation pipelines.

Skills

LLMs familiarity

ML models

Uncertainty concepts

Open-weight LLMs

Python on HPC

Software engineering

Independent work

Education

Bachelor’s degree or equivalent

Master’s degree in a related field

PhD student enrollment

Tools

Python

HPC systems

GPU clusters

Deep learning frameworks

Job description

The National Renewable Energy Laboratory is offering a graduate internship focused on quantifying uncertainty and reducing hallucinations in LLM-based science assistants. You will explore multi-turn, agentic workflows, assess internal representations, and develop detectors and probing methods, with experiments on HPC systems.

You will contribute to documentation, reports, and presentations while building robust research code and evaluation pipelines under a multidisciplinary team in Golden,